Question about battery test

[adam917]

Build is daily 2007-03-05. I'm using a repeated FLAC (Level album as my test material on the standard battery. Volume was set to 0 dB, backlight was set to off all the time on remote & main unit, and I used Sennheiser HD 555 headphones plugged into the remote as the audio output.

Is 14 h 23 min 37 s (from the battery bench log) an accurate representation for the stock 1300 mAh battery considering the settings and audio material I used? Do I have to run the test multiple times with even the same (loud) headphones in order to report it to the runtime page? I got this unit as 'new-old' stock off an eBay seller on 2006-04-30 (so it is a new unit, but who knows how long it's been since manufacture).

Also, is this test inaccurate as my settings are changed from the defaults?

Attached is my config.cfg that was used for the test with hardly anything modified from the way I usually use my unit and the bench log.

[attachment deleted by admin, too old]

[Llorean]

Honestly, tests should come in pairs:
One in Rockbox, and one in the original firmware using identical settings. This means you need to have settings in Rockbox you can reproduce in the original firmware, and shuffle should not be used, for example. This way you get to see whether your runtime is better or worse than the original firmware.

Unfortunately tests on their own don't really tell much, since you don't have a baseline for your battery condition. As well, formats like FLAC that decode faster than realtime, so don't require boosting, but involve spinning up the HD a lot, can't really be compared to lossy codecs like MP3, so it's hard to place that on a scale without other FLAC tests.

That being said, anyone is free to upload their results, but it's hard to compare them and know how things are doing in most cases.

[adam917]

So what would be an example of the best kind of benchmark? Leaving all settings on default and running a big enough MP3 album on repeat? Unfortunately, the device shut off when I was away. If I was doing this on the original firmware, I wouldn't know how long it lasted.

[adam917]

Honestly, tests should come in pairs:
One in Rockbox, and one in the original firmware using identical settings. This means you need to have settings in Rockbox you can reproduce in the original firmware, and shuffle should not be used, for example. This way you get to see whether your runtime is better or worse than the original firmware.

Unfortunately tests on their own don't really tell much, since you don't have a baseline for your battery condition. As well, formats like FLAC that decode faster than realtime, so don't require boosting, but involve spinning up the HD a lot, can't really be compared to lossy codecs like MP3, so it's hard to place that on a scale without other FLAC tests.

That being said, anyone is free to upload their results, but it's hard to compare them and know how things are doing in most cases.

So how do I upload my results? Do I have to go through the IRC procedure?

[Llorean]

Yes, you have to get write access to the wiki.